Video Individual Counting for Moving Drones

Video Individual Counting (VIC) has received increasing attention for its importance in intelligent video surveillance. Existing works are limited in two aspects, i.e., dataset and method. Previous datasets are captured with fixed or rarely moving cameras with relatively sparse individuals, restricting evaluation for a highly varying view and time in crowded scenes. Existing methods rely on localization fol lowed by association or classification, which struggle under dense and dynamic conditions due to inaccurate localiza tion of small targets. To address these issues, we introduce the MovingDroneCrowd Dataset, featuring videos captured by fast-moving drones in crowded scenes under diverse il luminations, shooting heights and angles. We further pro pose a Shared Density map-guided Network (SDNet) us ing a Depth-wise Cross-Frame Attention (DCFA) module to directly estimate shared density maps between consecu tive frames, from which the inflow and outflow density maps are derived by subtracting the shared density maps from the global density maps. The inflow density maps across frames are summed up to obtain the number of unique pedestrians in a video. Experiments on our datasets and publicly avail able ones show the superiority of our method over the state of the arts in highly dynamic and complex crowded scenes. Our dataset and codes have been released publicly (https://github.com/fyw1999/MovingDroneCrowd).

©2025 IEEE